INDEX

Explanations

the word "talk" and related words

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

emale

-0.76

boa

-0.69

arte

-0.68

iverpool

-0.65

uilt

-0.63

hews

-0.62

ritional

-0.61

rypt

-0.60

orius

-0.60

eele

-0.59

POSITIVE LOGITS

 about

0.99

 aloud

0.87

 louder

0.84

about

0.82

 ABOUT

0.82

 loudly

0.82

 smack

0.80

 frankly

0.77

ebus

0.75

ative

0.72

Activations Density 0.041%

the word "talk" and related words

No Comments

No Known Activations