INDEX

Explanations

abbreviations or acronyms followed by specific codes or terms within a structured format

oai_token-act-pair · gpt-3.5-turbo

references to various official acts, organizations, and systems

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Ļ½

-0.73

((

-0.68

:(

-0.68

oho

-0.67

(&

-0.67

:'

-0.65

angelo

-0.64

othe

-0.64

ãĥ´

-0.64

forth

-0.63

POSITIVE LOGITS

 Certification

0.65

 designation

0.63

 pione

0.62

 Committees

0.61

 moniker

0.60

 Regulations

0.59

Ltd

0.59

 acronym

0.58

system

0.58

/"

0.58

Activations Density 0.117%

abbreviations or acronyms followed by specific codes or terms within a structured format

references to various official acts, organizations, and systems

No Comments

No Known Activations