INDEX

Explanations

words related to news articles or reports

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_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

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

bilt

-0.78

 bearer

-0.71

«ĺ

-0.68

ICAN

-0.68

INAL

-0.66

ĻĤ

-0.65

 uncond

-0.62

 cracked

-0.61

 econom

-0.59

 privately

-0.59

POSITIVE LOGITS

uffs

0.85

ook

0.85

onga

0.81

nian

0.81

aru

0.77

amba

0.76

Forest

0.74

ornia

0.74

oo

0.73

igans

0.73

Activations Density 0.179%

words related to news articles or reports

No Comments

No Known Activations