INDEX

Explanations

information related to historical events or academic studies in different fields

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

iless

-0.64

ega

-0.63

LOG

-0.62

Â¶

-0.61

AG

-0.60

Voice

-0.59

SW

-0.59

eton

-0.58

twitch

-0.58

BAT

-0.57

POSITIVE LOGITS

 thereby

0.99

 consequently

0.98

how

0.97

its

0.96

 consequ

0.94

 whether

0.92

 thus

0.90

 hence

0.84

 subsequent

0.81

 surroundings

0.80

Activations Density 1.614%

information related to historical events or academic studies in different fields

No Comments

No Known Activations