INDEX

Explanations

phrases related to user-generated content and technical terms like 'mounted' and 'powered'

oai_token-act-pair · gpt-3.5-turbo

terms related to user-generated content and its various applications

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

KEY

-0.81

BIT

-0.77

natureconservancy

-0.71

phys

-0.70

adier

-0.69

alter

-0.67

forums

-0.66

codes

-0.66

itta

-0.66

osponsors

-0.65

POSITIVE LOGITS

nesday

0.81

nesses

0.77

ness

0.77

 subsequ

0.74

 versions

0.73

uled

0.73

theless

0.70

 suspensions

0.68

ãĤ´ãĥ³

0.67

terday

0.66

Activations Density 0.095%

phrases related to user-generated content and technical terms like 'mounted' and 'powered'

terms related to user-generated content and its various applications

No Comments

No Known Activations