INDEX

Explanations

words related to specific entities or proper nouns

oai_token-act-pair · gpt-3.5-turbo

proper nouns and technical terms

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

 âĢº

-0.53

 Sarah

-0.51

 Sonny

-0.51

¶ħ

-0.51

cmp

-0.51

ãĥīãĥ©ãĤ´ãĥ³

-0.50

ãĤ¸

-0.50

 Bryan

-0.49

 Kendrick

-0.49

 Gibbs

-0.48

POSITIVE LOGITS

intosh

0.57

 initiation

0.57

ulent

0.56

fu

0.54

Fil

0.54

icular

0.54

owship

0.54

utenberg

0.53

 aloud

0.53

ifer

0.51

Activations Density 0.742%

words related to specific entities or proper nouns

proper nouns and technical terms

No Comments

No Known Activations