INDEX

Explanations

mentions of statistical or numerical information

oai_token-act-pair · gpt-3.5-turbo

mentions of specific groups or entities, particularly numbers and statistics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

trump

-0.52

 guiActiveUn

-0.51

STATE

-0.50

 GOODMAN

-0.50

conom

-0.50

the

-0.49

querque

-0.47

mania

-0.47

avorite

-0.46

thel

-0.46

POSITIVE LOGITS

yn

0.47

To

0.45

occ

0.45

isy

0.43

ape

0.42

anse

0.42

Dew

0.42

 Forms

0.41

Mul

0.41

 Cath

0.41

Activations Density 0.570%

mentions of statistical or numerical information

mentions of specific groups or entities, particularly numbers and statistics

No Comments

No Known Activations