INDEX

Explanations

references to forensic and investigative activities

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

terms associated with investigative and forensic practices

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

zl

-0.85

yip

-0.83

lash

-0.78

eworld

-0.77

chery

-0.75

cair

-0.74

jri

-0.74

fty

-0.71

mun

-0.71

cens

-0.67

POSITIVE LOGITS

 Grimoire

0.89

 detectives

0.76

 detective

0.73

 trail

0.72

 investigator

0.71

sle

0.69

 interviewer

0.69

 reporters

0.69

 Bigfoot

0.69

 reporter

0.68

Activations Density 0.108%

references to forensic and investigative activities

terms associated with investigative and forensic practices

No Comments

No Known Activations