INDEX

Explanations

references to things that are described as generic

oai_token-act-pair · gpt-3.5-turbo

instances of the word "generic."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

oir

-0.74

kers

-0.74

otos

-0.73

=-=-=-=-=-=-=-=-

-0.73

cano

-0.72

Jenn

-0.70

 Ferry

-0.70

rey

-0.68

=-=-

-0.66

tein

-0.66

POSITIVE LOGITS

 generic

0.84

ization

0.81

isable

0.77

 interchange

0.75

 applic

0.74

ality

0.73

ALLY

0.71

generic

0.69

isation

0.69

ised

0.69

Activations Density 0.011%

references to things that are described as generic

instances of the word "generic."

No Comments

No Known Activations