INDEX

Explanations

text related to crime, arrests, and police incidents

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

culosis

-0.71

berus

-0.61

 Predators

-0.61

stasy

-0.59

 ceilings

-0.59

 briefs

-0.56

 calves

-0.55

iversal

-0.55

incinn

-0.55

 cumbers

-0.54

POSITIVE LOGITS

ridge

1.11

meal

0.97

aking

0.95

icularly

0.93

ridges

0.92

ners

0.89

icular

0.88

heartedly

0.85

ials

0.82

ially

0.82

Activations Density 1.191%

text related to crime, arrests, and police incidents

No Comments

No Known Activations