INDEX

Explanations

specific entities and locations, particularly related to law enforcement incidents and government activities

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

acronyms and initialisms related to agencies or organizations

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 12-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_post

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.11.hook_resid_post

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

tan

-0.65

gaard

-0.63

 imped

-0.60

âķ

-0.60

lessness

-0.59

len

-0.59

 Â¯

-0.58

phi

-0.57

tons

-0.56

ens

-0.55

POSITIVE LOGITS

senal

0.89

olitics

0.85

vantage

0.79

elaide

0.78

 Insight

0.73

ews

0.73

 News

0.72

today

0.71

 Radio

0.70

imes

0.68

Activations Density 0.136%

specific entities and locations, particularly related to law enforcement incidents and government activities

acronyms and initialisms related to agencies or organizations

No Comments

No Known Activations