INDEX

Explanations

phrases related to arguments and legal matters

oai_token-act-pair · gpt-3.5-turbo

specific job titles or roles within a context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

wcsstore

-0.72

 Chrom

-0.70

ãĥĦ

-0.69

artifacts

-0.67

 Jets

-0.64

^{

-0.63

 guiName

-0.62

erm

-0.62

gc

-0.61

Target

-0.61

POSITIVE LOGITS

vernment

0.84

idth

0.75

ector

0.72

 supplemented

0.70

 construed

0.69

ected

0.68

sembly

0.67

 heartbeat

0.62

 interpreted

0.61

izen

0.61

Activations Density 0.000%

phrases related to arguments and legal matters

specific job titles or roles within a context

No Comments

No Known Activations