INDEX

Explanations

proper nouns related to cultural or geographical references

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_fs12288-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

12,288

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ĳå£«

-0.66

tics

-0.66

suit

-0.66

 hazard

-0.65

APE

-0.63

deck

-0.62

ses

-0.62

tap

-0.62

sed

-0.61

side

-0.61

POSITIVE LOGITS

estro

1.40

ureen

1.31

ples

1.11

Ã±

1.02

uth

0.98

plin

0.96

arten

0.96

isel

0.94

pling

0.94

iami

0.91

Activations Density 0.048%

proper nouns related to cultural or geographical references

No Comments

No Known Activations