INDEX

Explanations

phrases expressing strong criticism or disbelief

oai_token-act-pair · gpt-3.5-turbo

instances of the word "nonsense" and related phrases

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

hani

-0.75

redits

-0.72

ez

-0.70

irth

-0.70

lis

-0.70

ugal

-0.69

hold

-0.69

uve

-0.69

yer

-0.68

imb

-0.65

POSITIVE LOGITS

 nonsense

1.11

 detector

0.91

 excuses

0.89

 bullshit

0.86

 excuse

0.83

 rubbish

0.81

 crap

0.79

 guiActiveUn

0.77

aceutical

0.77

 blah

0.77

Activations Density 0.027%

phrases expressing strong criticism or disbelief

instances of the word "nonsense" and related phrases

No Comments

No Known Activations