INDEX

Explanations

the specific number of occurrences of a particular action or event

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

Reviewer

-0.91

ists

-0.81

aceous

-0.78

XT

-0.77

DIT

-0.71

liction

-0.70

rats

-0.69

CHAT

-0.68

rera

-0.66

ourt

-0.66

POSITIVE LOGITS

 consecut

0.99

cale

0.81

pan

0.72

orial

0.70

points

0.69

 apiece

0.67

borough

0.63

coded

0.62

hare

0.62

fold

0.62

Activations Density 0.414%

the specific number of occurrences of a particular action or event

No Comments

No Known Activations