INDEX

Explanations

instances where something is only partially or partly done/completed

oai_token-act-pair · gpt-3.5-turbo

phrases indicating partial explanations or attributions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

ciating

-0.83

Trend

-0.83

eers

-0.81

Reviewer

-0.79

eer

-0.78

ãĥ¤

-0.73

eering

-0.71

mson

-0.70

yers

-0.70

pires

-0.70

POSITIVE LOGITS

 obscured

0.90

 cloudy

0.88

 paralyzed

0.86

 overlapping

0.78

 paraly

0.77

 compensate

0.77

 blame

0.76

 submerged

0.76

 blinded

0.76

 redacted

0.74

Activations Density 0.039%

instances where something is only partially or partly done/completed

phrases indicating partial explanations or attributions

No Comments

No Known Activations