INDEX

Explanations

self-referential phrases signaling the speaker's thoughts or actions

oai_token-act-pair · gpt-3.5-turbo

first-person perspective statements and expressions of personal thoughts or feelings

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

adra

-0.71

externalActionCode

-0.70

fig

-0.67

-0.64

×Ļ

-0.62

otiation

-0.60

WAR

-0.59

 Nanto

-0.59

edition

-0.59

 è£ıè¦ļéĨĴ

-0.58

POSITIVE LOGITS

 joking

0.90

 invincible

0.84

 kidding

0.77

 kindred

0.69

 innocuous

0.65

might

0.64

 might

0.62

amn

0.61

ļé

0.61

'd

0.61

Activations Density 0.169%

self-referential phrases signaling the speaker's thoughts or actions

first-person perspective statements and expressions of personal thoughts or feelings

No Comments

No Known Activations