INDEX

Explanations

phrases related to actions of taking something and transforming it into something else, typically in a negative or critical context

oai_token-act-pair · gpt-3.5-turbo

instances of the word "it" and related pronouns indicating transformation or conversion

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.10.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

accompan

-0.87

icipated

-0.74

erest

-0.67

rising

-0.66

 accompanies

-0.66

 Annotations

-0.64

è¦ļéĨĴ

-0.64

hess

-0.61

regate

-0.60

itely

-0.59

POSITIVE LOGITS

 into

1.26

into

1.20

 INTO

1.10

 Into

1.09

 upside

1.08

 inward

0.89

 sideways

0.86

wards

0.81

oward

0.79

 onto

0.78

Activations Density 0.078%

phrases related to actions of taking something and transforming it into something else, typically in a negative or critical context

instances of the word "it" and related pronouns indicating transformation or conversion

No Comments

No Known Activations