INDEX

Explanations

mentions of specific time periods or events

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

 Parenthood

-0.71

ufact

-0.68

edIn

-0.67

sbm

-0.67

fw

-0.66

avorite

-0.65

ropolitan

-0.64

renheit

-0.64

ategor

-0.64

anned

-0.61

POSITIVE LOGITS

 thereof

1.03

liest

0.95

waters

0.89

iest

0.86

of

0.84

most

0.80

side

0.76

adle

0.75

 zone

0.74

ultimate

0.72

Activations Density 1.237%

mentions of specific time periods or events

No Comments

No Known Activations