INDEX

Explanations

words related to political actions or intentions

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_fs12288-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

12,288

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

————

-0.73

ãĤ¡

-0.71

ettel

-0.68

ghan

-0.65

zyk

-0.65

gnu

-0.64

ibel

-0.63

endix

-0.63

schild

-0.61

atis

-0.61

POSITIVE LOGITS

hold

1.04

 sway

1.01

erness

0.99

 accountable

0.98

holders

0.97

holding

0.97

holder

0.92

 onto

0.90

overs

0.90

 hostage

0.86

Activations Density 0.083%

words related to political actions or intentions

No Comments

No Known Activations