INDEX

Explanations

phrases indicating outcomes or results

oai_token-act-pair · gpt-3.5-turbo

forms of the verb "to be" and related status descriptors

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

ourt

-0.64

inas

-0.64

EMBER

-0.64

cit

-0.62

volent

-0.62

iazep

-0.61

acist

-0.59

trop

-0.59

laughs

-0.59

conserv

-0.58

POSITIVE LOGITS

 therefore

0.93

 also

0.93

 however

0.82

 definitely

0.81

 certainly

0.79

 undoubtedly

0.77

 furthermore

0.75

 doubtless

0.72

 moreover

0.72

not

0.72

Activations Density 0.863%

phrases indicating outcomes or results

forms of the verb "to be" and related status descriptors

No Comments

No Known Activations