INDEX

Explanations

words related to consequences, actions, and results

oai_token-act-pair · gpt-3.5-turbo

terms related to legal actions and consequences

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

soever

-0.63

»Ĵ

-0.62

é¾įå¥ĳå£«

-0.62

atech

-0.58

©¶æ

-0.57

ITED

-0.56

orate

-0.56

rote

-0.55

handle

-0.55

 handle

-0.55

POSITIVE LOGITS

 downfall

0.89

 lasting

0.86

 termination

0.84

 paralysis

0.79

 eventual

0.76

 demise

0.74

 cancellation

0.72

 meltdown

0.71

 becoming

0.71

 escalation

0.70

Activations Density 0.408%

words related to consequences, actions, and results

terms related to legal actions and consequences

No Comments

No Known Activations