INDEX

Explanations

time-related phrases or expressions

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

omew

-0.66

ister

-0.64

omas

-0.61

clair

-0.61

 TAMADRA

-0.60

akening

-0.59

eday

-0.59

OTAL

-0.59

stal

-0.58

 Response

-0.58

POSITIVE LOGITS

 hoop

0.83

 fuss

0.81

 sudden

0.78

 bells

0.78

 stuff

0.78

 goodies

0.77

important

0.73

 facets

0.72

things

0.71

together

0.71

Activations Density 1.488%

time-related phrases or expressions

No Comments

No Known Activations