INDEX

Explanations

phrases related to expressing doubt or disbelief

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

entric

-0.84

ities

-0.74

orp

-0.72

risome

-0.70

 newsp

-0.69

 ILCS

-0.68

inational

-0.66

ciation

-0.66

iott

-0.64

pmwiki

-0.64

POSITIVE LOGITS

track

1.11

raise

0.83

coat

0.78

lash

0.77

stab

0.77

ped

0.77

dash

0.76

acked

0.75

side

0.75

packs

0.74

Activations Density 0.031%

phrases related to expressing doubt or disbelief

No Comments

No Known Activations