INDEX

Explanations

mentions of a specific entity or person named "Ner" with different numeric identifiers

oai_token-act-pair · gpt-3.5-turbo

proper nouns, particularly names of individuals

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

avorite

-0.71

carry

-0.68

UTC

-0.66

Constructed

-0.65

reaching

-0.64

EMP

-0.63

Breaking

-0.63

ERROR

-0.62

 runaway

-0.60

oral

-0.60

POSITIVE LOGITS

getic

1.02

ding

0.98

gie

0.96

ners

0.94

ning

0.90

lund

0.88

ner

0.86

nen

0.86

opol

0.86

amic

0.84

Activations Density 0.018%

mentions of a specific entity or person named "Ner" with different numeric identifiers

proper nouns, particularly names of individuals

No Comments

No Known Activations