INDEX

Explanations

proper nouns related to law enforcement and intelligence agencies

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

Interstitial

-0.75

SHIP

-0.74

 queens

-0.72

ð

-0.70

âĹ¼

-0.70

vana

-0.66

opian

-0.64

egal

-0.64

 sitcom

-0.64

ahime

-0.64

POSITIVE LOGITS

 Comey

1.12

FBI

0.97

 Clapper

0.94

FBI

0.90

 Mueller

0.88

 McCabe

0.86

 Brennan

0.82

 surveillance

0.76

ATF

0.76

 Sessions

0.76

Activations Density 0.099%

proper nouns related to law enforcement and intelligence agencies

No Comments

No Known Activations