INDEX

Explanations

names that contain the substring "ner"

oai_token-act-pair · gpt-3.5-turbo

the word "ner" in various contexts

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

-0.82

avorite

-0.66

UTC

-0.65

DonaldTrump

-0.65

EMP

-0.64

oral

-0.63

 Occupations

-0.62

ume

-0.62

wordpress

-0.60

orses

-0.60

POSITIVE LOGITS

ding

1.00

getic

0.97

dal

0.89

lund

0.88

ger

0.87

sonian

0.87

gie

0.86

jee

0.86

stein

0.85

ning

0.84

Activations Density 0.019%

names that contain the substring "ner"

the word "ner" in various contexts

No Comments

No Known Activations