INDEX

Explanations

the name "Gan" with varying intensities

oai_token-act-pair · gpt-3.5-turbo

occurrences of the name "Gan."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 differential

-0.75

 Jordanian

-0.63

 mates

-0.62

 isolation

-0.62

 entrants

-0.61

 contagious

-0.59

 sibling

-0.58

Department

-0.57

 admirable

-0.57

 outstanding

-0.57

POSITIVE LOGITS

gan

1.59

igans

1.25

culosis

1.03

ieri

0.95

igan

0.94

ciating

0.93

ecake

0.92

ghan

0.90

hesive

0.90

ijah

0.90

Activations Density 0.007%

the name "Gan" with varying intensities

occurrences of the name "Gan."

No Comments

No Known Activations