INDEX

Explanations

references to physical objects or concepts related to the word "plate."

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "plate."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

vernment

-1.05

rians

-0.78

Plug

-0.76

apest

-0.75

ITNESS

-0.72

ijah

-0.72

emouth

-0.71

uish

-0.69

Aud

-0.69

=]

-0.68

POSITIVE LOGITS

 plates

1.15

 plate

1.14

plate

1.06

meal

0.87

ographer

0.87

 Plate

0.84

washer

0.80

 lamps

0.80

ozo

0.78

 dishes

0.76

Activations Density 0.012%

references to physical objects or concepts related to the word "plate."

occurrences of the word "plate."

No Comments

No Known Activations