INDEX

Explanations

numerical sequences and data-related information, as well as mentions of software tools and commands

oai_token-act-pair · gpt-3.5-turbo

numerical data and references to time or dates

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

gmail

-0.72

aird

-0.71

.")

-0.62

cause

-0.61

gross

-0.59

sic

-0.59

 besides

-0.58

inence

-0.58

isky

-0.58

inant

-0.57

POSITIVE LOGITS

--+

0.72

 Awakens

0.69

[*]

0.68

ãĢĳ

0.67

©¶æ

0.67

ãĥ³ãĤ¸

0.67

 Bayern

0.66

 Dortmund

0.63

DLC

0.62

eca

0.61

Activations Density 0.348%

numerical sequences and data-related information, as well as mentions of software tools and commands

numerical data and references to time or dates

No Comments

No Known Activations