INDEX

Explanations

names of individuals

oai_token-act-pair · gpt-3.5-turbo

proper nouns, particularly names and possibly locations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 EDITION

-0.77

 STATS

-0.67

 Presidents

-0.66

 Rolls

-0.65

-0.62

mpire

-0.62

NYSE

-0.62

 arenas

-0.60

PRES

-0.60

GST

-0.59

POSITIVE LOGITS

irez

0.90

 recalled

0.81

 disappeared

0.77

Doe

0.76

 Tsarnaev

0.74

alez

0.74

inson

0.73

 abducted

0.73

essa

0.72

yk

0.71

Activations Density 0.281%

names of individuals

proper nouns, particularly names and possibly locations

No Comments

No Known Activations