INDEX

Explanations

political terms and news-related keywords

oai_token-act-pair · gpt-3.5-turbo

references to political emergency declarations and related contextual information

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

fal

-0.58

 stall

-0.57

GW

-0.57

ox

-0.56

Neal

-0.56

IRO

-0.55

IPM

-0.55

MI

-0.54

Reloaded

-0.54

Bul

-0.54

POSITIVE LOGITS

 respectively

1.03

aughs

0.74

ilitary

0.73

ovies

0.71

.).

0.71

quet

0.70

ÃĥÃĤ

0.70

culated

0.69

qqa

0.68

igious

0.68

Activations Density 1.848%

political terms and news-related keywords

references to political emergency declarations and related contextual information

No Comments

No Known Activations