INDEX

Explanations

phrases related to legal proceedings and law enforcement actions

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

 undertaking

-0.68

Hai

-0.61

 Sandwich

-0.60

 Barg

-0.60

 Mysteries

-0.60

dom

-0.59

 issu

-0.59

hood

-0.58

nesty

-0.58

 achie

-0.58

POSITIVE LOGITS

fully

1.25

FUL

1.13

fulness

0.99

 sparing

0.92

ful

0.92

 condoms

0.88

full

0.84

 aliases

0.75

 techniques

0.73

ibles

0.72

Activations Density 0.629%

phrases related to legal proceedings and law enforcement actions

No Comments

No Known Activations