INDEX

Explanations

pronouns or nouns indicating people

oai_token-act-pair · gpt-3.5-turbo

pronouns related to collective or individual experiences and identities

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

————

-0.69

 Anat

-0.67

Correction

-0.66

 Canaver

-0.65

Interested

-0.62

 Bundes

-0.61

RAY

-0.59

 Britt

-0.59

 Unlimited

-0.59

-0.58

POSITIVE LOGITS

©¶æ

0.82

 encount

0.80

ngth

0.80

cius

0.76

 frequ

0.75

ÃĥÃĤ

0.74

 attained

0.72

 encountered

0.70

alian

0.70

 consume

0.70

Activations Density 0.325%

pronouns or nouns indicating people

pronouns related to collective or individual experiences and identities

No Comments

No Known Activations