INDEX

Explanations

phrases related to mistreatment or misidentification

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "mist" and its variations in different contexts

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

jriwal

-0.84

kefeller

-0.72

thodox

-0.71

lishes

-0.71

Books

-0.69

 Apex

-0.66

audi

-0.66

illac

-0.65

MIT

-0.65

adesh

-0.65

POSITIVE LOGITS

reatment

1.27

reated

1.12

rans

0.97

reating

0.94

arget

0.89

aken

0.87

akable

0.86

ypes

0.86

 mist

0.83

mallow

0.83

Activations Density 0.007%

phrases related to mistreatment or misidentification

occurrences of the word "mist" and its variations in different contexts

No Comments

No Known Activations