INDEX

Explanations

names mentioned in a text

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "name" as part of personal introductions

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

yrinth

-0.92

EMS

-0.92

aunders

-0.88

Js

-0.80

idth

-0.74

romy

-0.72

erg

-0.71

ourse

-0.71

BSD

-0.70

iths

-0.70

POSITIVE LOGITS

plates

1.03

plate

0.96

 aliases

0.93

tag

0.87

paces

0.86

 tags

0.83

 surname

0.81

 redacted

0.78

akes

0.77

 surn

0.74

Activations Density 0.029%

names mentioned in a text

occurrences of the word "name" as part of personal introductions

No Comments

No Known Activations