INDEX

Explanations

proper nouns, specifically names of locations, organizations, and individuals

oai_token-act-pair · gpt-3.5-turbo

references to specific political parties or movements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

Wall

-0.67

ãĤ¯

-0.67

SHARE

-0.66

Bear

-0.66

strong

-0.66

Strength

-0.66

protected

-0.66

 Lieberman

-0.65

Wide

-0.65

options

-0.62

POSITIVE LOGITS

eers

0.87

entimes

0.78

bourg

0.76

 charm

0.76

ernel

0.69

)]

0.68

 aspir

0.67

 faculties

0.66

ciating

0.65

olitics

0.64

Activations Density 0.090%

proper nouns, specifically names of locations, organizations, and individuals

references to specific political parties or movements

No Comments

No Known Activations