INDEX

Explanations

instances where the number "one" is emphasized or highlighted for specific actions or items

oai_token-act-pair · gpt-3.5-turbo

references to the word "one" 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

ooks

-0.76

ãĤµ

-0.64

inders

-0.63

æĸ¹

-0.62

folk

-0.60

lov

-0.59

hips

-0.59

 actionGroup

-0.59

inas

-0.58

ãĤ¬

-0.58

POSITIVE LOGITS

 Hundred

0.98

 hundred

0.98

rency

0.85

 dimensional

0.85

 thousand

0.83

 Thousand

0.77

anchester

0.75

handedly

0.75

 sided

0.73

 million

0.72

Activations Density 0.180%

instances where the number "one" is emphasized or highlighted for specific actions or items

references to the word "one" in various contexts

No Comments

No Known Activations