INDEX

Explanations

comparative adjectives ending in '-lier' or '-liest'

oai_token-act-pair · gpt-3.5-turbo

words related to being an outlier or unusual characteristics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

raham

-0.74

issued

-0.71

UTC

-0.71

ushed

-0.69

azel

-0.68

elin

-0.68

href

-0.68

keyes

-0.67

shire

-0.67

oling

-0.66

POSITIVE LOGITS

lihood

0.98

 contingency

0.73

 estim

0.66

 Vict

0.64

 Than

0.64

 Spir

0.62

 guiName

0.62

 pokemon

0.61

 occurrences

0.61

 prey

0.60

Activations Density 0.021%

comparative adjectives ending in '-lier' or '-liest'

words related to being an outlier or unusual characteristics

No Comments

No Known Activations