INDEX

Explanations

words or phrases related to interception or intercepting

oai_token-act-pair · gpt-3.5-turbo

terms related to interception and monitoring activities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Chau

-0.75

abe

-0.70

 pains

-0.70

 Creative

-0.70

uer

-0.69

chair

-0.68

NOW

-0.67

Gau

-0.67

 Muse

-0.65

obs

-0.64

POSITIVE LOGITS

 intercept

4.01

 intercepted

2.60

 Intercept

2.26

 interception

1.89

 interceptions

1.40

 eaves

1.37

 decrypt

1.26

 wiret

1.23

 scram

1.18

 deflect

1.04

Activations Density 0.018%

words or phrases related to interception or intercepting

terms related to interception and monitoring activities

No Comments

No Known Activations