INDEX

Explanations

references to geographical entities, specifically provinces

oai_token-act-pair · gpt-3.5-turbo

references to a specific province

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

visory

-0.81

kus

-0.79

noon

-0.76

artz

-0.76

tle

-0.75

eor

-0.75

ammad

-0.73

bugs

-0.72

paces

-0.72

chel

-0.72

POSITIVE LOGITS

 province

1.07

 provinces

0.88

 sovere

0.87

 unemploy

0.82

 Province

0.81

 annexed

0.78

 provincial

0.78

 rall

0.77

wide

0.77

 constitu

0.77

Activations Density 0.011%

references to geographical entities, specifically provinces

references to a specific province

No Comments

No Known Activations