INDEX

Explanations

phrases related to frequency or routine, especially with a focus on medical or technical contexts

oai_token-act-pair · gpt-3.5-turbo

phrases describing regular or recurring activities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 Patriarch

-0.82

enic

-0.72

Els

-0.69

 Gree

-0.67

atories

-0.67

 Torn

-0.66

agra

-0.66

 Sins

-0.65

erness

-0.64

 Stall

-0.64

POSITIVE LOGITS

 basis

1.08

 thereafter

0.81

iatus

0.81

uation

0.75

 continuum

0.72

cale

0.71

 standpoint

0.68

forward

0.66

VERTISEMENT

0.64

mma

0.64

Activations Density 0.030%

phrases related to frequency or routine, especially with a focus on medical or technical contexts

phrases describing regular or recurring activities

No Comments

No Known Activations