INDEX

Explanations

words related to taking action or making something happen

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

stration

-0.62

thia

-0.61

;;;;

-0.61

 presided

-0.59

 alluded

-0.59

rous

-0.59

 Niet

-0.58

hood

-0.57

ander

-0.57

below

-0.56

POSITIVE LOGITS

 sure

1.35

hift

1.04

 adjustments

0.88

 sense

0.84

URE

0.76

 decisions

0.76

 excuses

0.75

 repairs

0.74

solete

0.73

 Sense

0.73

Activations Density 0.105%

words related to taking action or making something happen

No Comments

No Known Activations