INDEX

Explanations

expressions related to caution or warnings

oai_token-act-pair · gpt-3.5-turbo

expressions of caution and disclaimers in opinionated statements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

Gap

-0.59

eele

-0.59

 potent

-0.57

reens

-0.57

 intercepted

-0.56

ortment

-0.56

 vanished

-0.56

Almost

-0.55

ushi

-0.55

 Layer

-0.54

POSITIVE LOGITS

 myself

1.21

nor

0.95

 anymore

0.94

 specifics

0.93

 condone

0.85

elo

0.81

 personally

0.79

 anybody

0.75

bole

0.72

yet

0.70

Activations Density 0.323%

expressions related to caution or warnings

expressions of caution and disclaimers in opinionated statements

No Comments

No Known Activations