INDEX

Explanations

geographical locations and specific details mentioned in news articles

oai_token-act-pair · gpt-3.5-turbo

geographical locations and references to entities, such as businesses or educational institutions

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

 unlike

-0.54

 Siber

-0.52

 disabling

-0.50

DERR

-0.49

 Berk

-0.48

 Pengu

-0.47

ãĤŃ

-0.47

pi

-0.46

 differently

-0.46

 Kafka

-0.44

POSITIVE LOGITS

 appeared

0.93

 ensured

0.90

 succumbed

0.86

 surfaced

0.84

 emerged

0.84

 collided

0.83

 prevailed

0.83

 appears

0.83

 began

0.82

 resulted

0.81

Activations Density 0.880%

geographical locations and specific details mentioned in news articles

geographical locations and references to entities, such as businesses or educational institutions

No Comments

No Known Activations