INDEX

Explanations

instances of rankings or positions in lists

oai_token-act-pair · gpt-3.5-turbo

quantitative rankings or statistical positions of entities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

uala

-0.77

issue

-0.72

unity

-0.71

Ã§a

-0.70

ĸļ

-0.68

Clar

-0.67

actionDate

-0.67

elcome

-0.67

quant

-0.67

ãĤ¶

-0.66

POSITIVE LOGITS

 earners

1.09

 lists

1.06

%:

1.04

 contenders

1.02

 performers

1.00

 finishes

0.98

 list

0.96

 spots

0.95

 finalists

0.95

 rankings

0.95

Activations Density 0.092%

instances of rankings or positions in lists

quantitative rankings or statistical positions of entities

No Comments

No Known Activations