INDEX

Explanations

items that are sequential in a list

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

ties

-0.71

sed

-0.69

vim

-0.69

bos

-0.66

walking

-0.65

sav

-0.64

pher

-0.63

 Gould

-0.62

dies

-0.61

nz

-0.61

POSITIVE LOGITS

 responders

1.08

 impressions

0.90

 baseman

0.85

 Steps

0.78

 glance

0.77

 ancest

0.75

terday

0.74

 Respond

0.72

 Corinthians

0.71

 Thoughts

0.69

Activations Density 0.057%

items that are sequential in a list

No Comments

No Known Activations