INDEX

Explanations

words related to impacts or interactions, with a particular focus on actions or events having a strong effect or outcome

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

pires

-0.66

ç«

-0.64

href

-0.63

Loading

-0.63

ais

-0.62

otype

-0.60

atu

-0.60

åŃ

-0.59

agin

-0.59

æĥ

-0.59

POSITIVE LOGITS

ched

1.11

achi

0.88

boxes

0.87

ches

0.84

ting

0.81

ted

0.81

waves

0.79

 hardest

0.79

pell

0.78

 puberty

0.76

Activations Density 2.362%

words related to impacts or interactions, with a particular focus on actions or events having a strong effect or outcome

No Comments

No Known Activations