INDEX

Explanations

phrases related to discussions, particularly about responsibility and blaming

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

ONSORED

-0.89

flies

-0.75

SourceFile

-0.75

ACT

-0.73

waters

-0.69

MODE

-0.68

reference

-0.68

YY

-0.68

lege

-0.68

©¶æ¥µ

-0.68

POSITIVE LOGITS

 behalf

1.79

erous

1.50

shore

0.99

 arrival

0.91

etime

0.84

slaught

0.84

 Saturdays

0.82

 Sundays

0.82

 steroids

0.80

site

0.78

Activations Density 0.606%

phrases related to discussions, particularly about responsibility and blaming

No Comments

No Known Activations