INDEX

Explanations

highly intensified or exaggerated words

oai_token-act-pair · gpt-3.5-turbo

phrases related to heightened emotions or states of hyperactivity

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

ORED

-0.73

Jaw

-0.71

 Geneva

-0.69

oured

-0.68

 goodbye

-0.68

DERR

-0.68

 Forever

-0.67

 Jinn

-0.66

OME

-0.63

 wiser

-0.62

POSITIVE LOGITS

bole

1.58

visor

1.43

visors

1.23

dimension

1.20

bol

1.16

vent

1.15

links

1.09

tro

1.08

active

1.07

dimensional

1.05

Activations Density 0.013%

highly intensified or exaggerated words

phrases related to heightened emotions or states of hyperactivity

No Comments

No Known Activations