INDEX

Explanations

phrases introducing new information or explaining a situation

oai_token-act-pair · gpt-3.5-turbo

repetitive conjunctions used to link ideas

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

ammy

-0.70

sha

-0.64

Ranked

-0.60

 laun

-0.60

intosh

-0.58

degree

-0.57

ICT

-0.56

mile

-0.56

ROR

-0.54

chieve

-0.54

POSITIVE LOGITS

oner

1.07

bered

1.04

othe

1.00

 congr

0.99

apy

0.96

FTWARE

0.95

ooo

0.93

oooo

0.86

 hopefully

0.86

 yeah

0.85

Activations Density 0.079%

phrases introducing new information or explaining a situation

repetitive conjunctions used to link ideas

No Comments

No Known Activations