INDEX

Explanations

references to subway stations

oai_token-act-pair · gpt-3.5-turbo

mentions of subway systems

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

wine

-0.74

iyah

-0.72

cius

-0.71

ificial

-0.71

erenn

-0.69

abama

-0.69

ivia

-0.68

laus

-0.67

 Christensen

-0.66

arget

-0.66

POSITIVE LOGITS

 subway

1.08

 Subway

0.98

 commute

0.89

 station

0.84

 commuting

0.83

 stations

0.81

 trains

0.81

 Transit

0.79

roads

0.78

 entrances

0.78

Activations Density 0.019%

references to subway stations

mentions of subway systems

No Comments

No Known Activations