INDEX

Explanations

sharp objects or instances described as sharp

oai_token-act-pair · gpt-3.5-turbo

references to sharp objects or sharpness in general

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

mits

-0.79

pty

-0.79

ople

-0.74

rella

-0.72

ORED

-0.70

ccess

-0.70

BLE

-0.69

mitting

-0.69

uthor

-0.66

avis

-0.66

POSITIVE LOGITS

ened

1.07

 sharp

1.03

ening

0.95

 sharper

0.87

ness

0.85

sharp

0.82

 blade

0.81

 distinction

0.80

ener

0.76

 spikes

0.76

Activations Density 0.009%

sharp objects or instances described as sharp

references to sharp objects or sharpness in general

No Comments

No Known Activations