INDEX

Explanations

mentions of radio stations

oai_token-act-pair · gpt-3.5-turbo

references to radio and broadcasting stations

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

stract

-0.73

yles

-0.71

llah

-0.68

jri

-0.67

 Sachs

-0.67

xtap

-0.67

leans

-0.67

ggie

-0.67

agna

-0.67

illard

-0.66

POSITIVE LOGITS

 broadcasts

0.98

 broadcaster

0.93

 broadcasting

0.92

 broadcasters

0.88

 Broadcasting

0.87

 announcer

0.85

 broadcast

0.84

 aired

0.82

 transmitter

0.82

 stations

0.80

Activations Density 0.026%

mentions of radio stations

references to radio and broadcasting stations

No Comments

No Known Activations