INDEX

Explanations

mentions of specific numbers or quantitative data

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

osi

-0.69

 Crown

-0.66

anding

-0.66

anwhile

-0.61

POS

-0.55

entary

-0.54

wait

-0.54

ocracy

-0.54

dding

-0.54

osa

-0.54

POSITIVE LOGITS

 gotten

1.25

 been

1.23

gotten

1.09

been

1.04

 gone

1.01

 Been

0.93

 undergone

0.88

 underestimated

0.87

 begun

0.87

 slipped

0.87

Activations Density 0.046%

mentions of specific numbers or quantitative data

No Comments

No Known Activations