INDEX

Explanations

phrases where something is identified as part of a larger whole or system

oai_token-act-pair · gpt-3.5-turbo

instances of the word "part" in relation to various topics or concepts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

ebted

-0.68

 candles

-0.67

 viol

-0.64

ancies

-0.64

 ceilings

-0.64

 aside

-0.62

 sisters

-0.60

oa

-0.60

 notwithstanding

-0.60

 compares

-0.59

POSITIVE LOGITS

axy

0.69

 everyday

0.69

 broader

0.68

the

0.67

amaru

0.67

erville

0.65

scription

0.65

 folklore

0.65

ortment

0.65

ĺħ

0.64

Activations Density 0.099%

phrases where something is identified as part of a larger whole or system

instances of the word "part" in relation to various topics or concepts

No Comments

No Known Activations