INDEX

Explanations

specific references to Formula 1 car numbers

oai_token-act-pair · gpt-3.5-turbo

references to Formula 1 racing and its associated terms

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

onse

-0.79

olulu

-0.78

WARD

-0.77

ages

-0.75

eston

-0.75

raviolet

-0.73

ritic

-0.73

aspberry

-0.71

thodox

-0.69

orative

-0.68

POSITIVE LOGITS

 suprem

0.81

 Classics

0.81

 Racing

0.78

 racing

0.75

 Clubs

0.74

 Sport

0.74

 legend

0.74

 racer

0.73

 championship

0.73

 Championships

0.73

Activations Density 0.034%

specific references to Formula 1 car numbers

references to Formula 1 racing and its associated terms

No Comments

No Known Activations