INDEX

Explanations

specifications or references related to popular films and franchises

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs49152-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

Data Type

torch.float32

Hook Point

blocks.8.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

 elev

-0.66

suc

-0.65

ĪĴ

-0.65

 peanuts

-0.62

 knees

-0.62

 sooner

-0.62

 illegitimate

-0.61

_>

-0.60

oler

-0.60

 cursing

-0.60

POSITIVE LOGITS

 Courtesy

0.96

 Â©

0.95

Courtesy

0.88

 Provided

0.85

 COURT

0.85

 Photograph

0.80

 Bett

0.79

 Getty

0.79

 REUTERS

0.77

 Contemporary

0.75

Activations Density 0.015%

specifications or references related to popular films and franchises

No Comments

No Known Activations