INDEX

Explanations

proper nouns, specifically names of people and places

oai_token-act-pair · gpt-3.5-turbo

proper nouns and specific names associated with significant locations or individuals

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

gements

-0.97

gments

-0.96

tymology

-0.88

 corrid

-0.82

Ð¾

-0.82

tarians

-0.80

tarian

-0.79

ÑĢ

-0.79

oiler

-0.77

pse

-0.77

POSITIVE LOGITS

cca

0.94

Cou

0.82

 Cabrera

0.78

 McMahon

0.75

Cue

0.74

 Mecca

0.72

 Heard

0.72

 Scotia

0.72

icago

0.67

Cur

0.66

Activations Density 0.047%

proper nouns, specifically names of people and places

proper nouns and specific names associated with significant locations or individuals

No Comments

No Known Activations