INDEX

Explanations

words related to interception or intercepting

oai_token-act-pair · gpt-3.5-turbo

terms related to interception and surveillance activities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

pet

-0.71

cil

-0.69

bold

-0.67

adders

-0.64

giving

-0.64

Year

-0.63

Myth

-0.63

Metro

-0.63

Sort

-0.62

burning

-0.61

POSITIVE LOGITS

 intercept

1.33

 intercepted

1.04

ãĥ¼ãĥĨãĤ£

0.95

rons

0.91

inel

0.89

ially

0.84

ibles

0.83

ezvous

0.82

uality

0.82

awei

0.81

Activations Density 0.007%

words related to interception or intercepting

terms related to interception and surveillance activities

No Comments

No Known Activations