INDEX

Explanations

names, possibly for identification or profiling purposes

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 crocod

-0.71

ting

-0.70

gression

-0.64

lli

-0.63

Charge

-0.63

 DevOnline

-0.61

tub

-0.61

ãĥ³ãĤ¸

-0.59

table

-0.59

 Perception

-0.59

POSITIVE LOGITS

elong

1.12

essee

1.08

esse

1.03

etts

0.97

etry

0.96

etary

0.96

ety

0.95

eteen

0.94

sylvania

0.94

arding

0.93

Activations Density 0.036%

names, possibly for identification or profiling purposes

No Comments

No Known Activations