INDEX

Explanations

mentions of locations or institutions

oai_token-act-pair · gpt-3.5-turbo

references to authoritative sources in a text

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

ãĥĺ

-0.78

obyl

-0.78

NetMessage

-0.77

adow

-0.74

icably

-0.73

hyde

-0.72

lessness

-0.69

Slot

-0.67

azes

-0.67

needed

-0.66

POSITIVE LOGITS

 latest

1.18

 report

1.11

 affidavit

1.03

 website

1.01

 statement

1.00

 biography

0.99

 Associated

0.98

 Huffington

0.98

 Encyclopedia

0.93

 Daily

0.93

Activations Density 0.119%

mentions of locations or institutions

references to authoritative sources in a text

No Comments

No Known Activations