INDEX

Explanations

words related to sharp objects like 'needle' and 'compass'

oai_token-act-pair · gpt-3.5-turbo

references to needles in a medical or metaphorical context

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

edom

-0.78

onym

-0.71

ARCH

-0.70

estation

-0.70

 Democr

-0.70

ership

-0.69

ivia

-0.67

*/(

-0.67

VERTISEMENT

-0.66

orial

-0.64

POSITIVE LOGITS

 needle

1.50

 needles

1.35

backer

1.17

 stitching

0.86

hook

0.82

 flared

0.80

 filament

0.78

ringe

0.78

 cartridges

0.75

 flare

0.74

Activations Density 0.009%

words related to sharp objects like 'needle' and 'compass'

references to needles in a medical or metaphorical context

No Comments

No Known Activations