INDEX

Explanations

information related to political figures, committees, investigations, and governmental actions

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

Characters

-0.71

escal

-0.69

antics

-0.66

 VIDEOS

-0.65

ģ«

-0.65

achy

-0.65

 giveaways

-0.64

rums

-0.63

 warranted

-0.63

isms

-0.60

POSITIVE LOGITS

 grandson

0.99

 pseudonym

0.95

 longtime

0.94

 former

0.94

 founder

0.91

 famed

0.89

 renowned

0.89

 nephew

0.88

 retired

0.88

 founding

0.85

Activations Density 1.884%

information related to political figures, committees, investigations, and governmental actions

No Comments

No Known Activations