INDEX

Explanations

the prefix "con-" in words

oai_token-act-pair · gpt-3.5-turbo

instances of the prefix "con," indicating concepts of connection, containment, or containment-related terminology

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

TPS

-0.76

HAEL

-0.72

ASED

-0.71

 Memories

-0.67

 Schwar

-0.67

ARD

-0.66

DIR

-0.63

Lur

-0.63

 Maple

-0.63

POSITIVE LOGITS

clave

1.09

clusively

1.02

gru

0.99

nex

0.93

verse

0.91

notation

0.90

formance

0.89

clus

0.88

ventions

0.87

con

0.86

Activations Density 0.006%

the prefix "con-" in words

instances of the prefix "con," indicating concepts of connection, containment, or containment-related terminology

No Comments

No Known Activations