INDEX

Explanations

mentions of the presence or continuation of certain states or actions over time

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

insula

-0.72

 Stim

-0.69

 elimination

-0.65

iHUD

-0.61

stice

-0.60

alties

-0.59

 diversion

-0.59

ongyang

-0.59

é¾

-0.57

 Moody

-0.57

POSITIVE LOGITS

birth

1.14

 intact

0.93

born

0.89

 unanswered

0.88

 undecided

0.78

 reeling

0.78

 alive

0.77

atile

0.74

 extant

0.72

 Nadu

0.72

Activations Density 0.978%

mentions of the presence or continuation of certain states or actions over time

No Comments

No Known Activations