INDEX

Explanations

terms related to interception, such as intercept, interceptor, and interceptions

oai_token-act-pair · gpt-3.5-turbo

mentions of the term "Intercept" related to surveillance or military contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 deaf

-0.80

WAYS

-0.65

UGE

-0.65

oke

-0.64

Downloadha

-0.63

creen

-0.63

HAHA

-0.61

\\\\

-0.61

cker

-0.61

HAHAHAHA

-0.59

POSITIVE LOGITS

 Intercept

1.08

ors

1.06

abad

0.94

rador

0.93

rained

0.90

rons

0.90

aired

0.86

oway

0.80

ions

0.79

illian

0.76

Activations Density 0.012%

terms related to interception, such as intercept, interceptor, and interceptions

mentions of the term "Intercept" related to surveillance or military contexts

No Comments

No Known Activations