INDEX

Explanations

words related to famous individuals

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

schild

-0.66

ÄŁ

-0.65

enegger

-0.61

 Cummings

-0.58

ynamic

-0.58

usercontent

-0.57

 Brill

-0.56

ktop

-0.56

 Daly

-0.56

WAY

-0.56

POSITIVE LOGITS

ongyang

0.95

sylvania

0.91

ciation

0.86

ĵĺ

0.82

asus

0.81

umatic

0.78

uese

0.72

formance

0.72

lator

0.71

Cola

0.70

Activations Density 2.582%

words related to famous individuals

No Comments

No Known Activations