INDEX

Explanations

references to evidence or proof

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

ategory

-0.81

Hop

-0.73

ttle

-0.73

aeper

-0.71

ernel

-0.70

scill

-0.68

iery

-0.68

otom

-0.65

 Chop

-0.65

 throats

-0.64

POSITIVE LOGITS

evidence

1.06

 evidence

1.02

 tampering

0.96

 corrobor

0.93

 linking

0.91

 indicating

0.90

Evidence

0.90

 evid

0.87

 demonstrating

0.86

 suggesting

0.86

Activations Density 0.537%

references to evidence or proof

No Comments

No Known Activations