INDEX

Explanations

references to drives or driving, such as physical drives or motivational drives

oai_token-act-pair · gpt-3.5-turbo

instances of the word "drive" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 Seym

-1.06

umbn

-0.73

ileaks

-0.71

ANN

-0.69

ereo

-0.69

 Lumpur

-0.69

iannopoulos

-0.67

aido

-0.67

gd

-0.65

ikhail

-0.64

POSITIVE LOGITS

bike

0.98

wheel

0.92

 drives

0.90

away

0.88

drive

0.85

train

0.85

 drive

0.84

 driving

0.77

boats

0.75

whe

0.74

Activations Density 0.020%

references to drives or driving, such as physical drives or motivational drives

instances of the word "drive" in various contexts

No Comments

No Known Activations