INDEX

Explanations

words related to updates, news, and progress

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

onz

-0.66

bilt

-0.63

sted

-0.62

oso

-0.61

hedon

-0.60

 chains

-0.60

iard

-0.59

sha

-0.58

toe

-0.58

 corps

-0.58

POSITIVE LOGITS

 happening

1.28

 transpired

1.04

 happ

0.99

 happened

0.95

 occurring

0.84

 happen

0.81

 happens

0.79

 done

0.78

 Happ

0.76

 bothering

0.74

Activations Density 0.586%

words related to updates, news, and progress

No Comments

No Known Activations