INDEX

Explanations

phrases or words related to a person's name, specifically with the string "ns" followed by a number

oai_token-act-pair · gpt-3.5-turbo

instances of the word "hens."

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

PDATE

-0.79

OSS

-0.68

MENTS

-0.65

 conclud

-0.63

BIL

-0.63

Desk

-0.62

SIGN

-0.60

RANT

-0.60

 Galile

-0.60

 Americas

-0.59

POSITIVE LOGITS

uits

1.14

uit

1.08

hers

1.06

pire

1.01

etsk

0.97

pie

0.96

mas

0.96

cript

0.95

pring

0.93

hao

0.92

Activations Density 0.023%

phrases or words related to a person's name, specifically with the string "ns" followed by a number

instances of the word "hens."

No Comments

No Known Activations