INDEX

Explanations

sequences featuring the word "then" followed by different situations or actions

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_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

toe

-0.78

orum

-0.66

than

-0.66

md

-0.65

aez

-0.62

cos

-0.61

scription

-0.60

caps

-0.60

chio

-0.59

 belief

-0.58

POSITIVE LOGITS

 proceeded

1.10

 proceed

0.88

EStream

0.78

 proceeds

0.78

igslist

0.77

 promptly

0.75

 secondly

0.74

 suddenly

0.74

Ń·

0.74

 abruptly

0.73

Activations Density 0.411%

sequences featuring the word "then" followed by different situations or actions

No Comments

No Known Activations