INDEX

Explanations

words related to intelligence, understanding, or investigation

oai_token-act-pair · gpt-3.5-turbo

instances of the substring "int."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 disbelief

-0.68

 Mbps

-0.67

ggle

-0.64

prototype

-0.60

 Generation

-0.60

 heaven

-0.56

 toes

-0.56

 smartphones

-0.56

 unprepared

-0.55

Accessory

-0.55

POSITIVE LOGITS

itial

0.90

rep

0.89

repre

0.82

edi

0.82

icter

0.81

ros

0.81

ented

0.79

aign

0.77

ruct

0.75

atra

0.74

Activations Density 0.040%

words related to intelligence, understanding, or investigation

instances of the substring "int."

No Comments

No Known Activations