INDEX

Explanations

words related to underground or hidden activities

oai_token-act-pair · gpt-3.5-turbo

instances of the word "under" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 testify

-0.70

ewitness

-0.66

ioxide

-0.65

bernatorial

-0.64

ichita

-0.63

 sidx

-0.61

IER

-0.60

 testified

-0.60

-0.59

 Madrid

-0.59

POSITIVE LOGITS

neath

1.09

standing

1.05

stood

0.99

wear

0.98

lings

0.95

blind

0.94

under

0.90

graduate

0.89

pins

0.87

tain

0.86

Activations Density 0.008%

words related to underground or hidden activities

instances of the word "under" in various contexts

No Comments

No Known Activations