INDEX

Explanations

proper names or abbreviations with the pattern P. <initial><surname>

oai_token-act-pair · gpt-3.5-turbo

references to specific individuals, particularly those with the initial 'P.'

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.5.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

ĸļ

-0.84

ources

-0.66

schild

-0.66

enhagen

-0.59

ailability

-0.58

 Spiegel

-0.57

 Normandy

-0.55

azeera

-0.54

ngth

-0.53

WAY

-0.53

POSITIVE LOGITS

jee

0.75

ople

0.59

twitter

0.59

ĵĺ

0.59

nels

0.58

 ACTIONS

0.57

sylvania

0.57

itent

0.56

asus

0.55

ciation

0.55

Activations Density 0.913%

proper names or abbreviations with the pattern P. <initial><surname>

references to specific individuals, particularly those with the initial 'P.'

No Comments

No Known Activations