INDEX

Explanations

words related to a specific name, potentially "Damon."

oai_token-act-pair · gpt-3.5-turbo

references to "ors," possibly indicating comparisons or alternatives 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 @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

Fra

-0.66

Sus

-0.65

à¥

-0.62

OTAL

-0.62

URRENT

-0.61

SIM

-0.61

shell

-0.60

Wall

-0.60

-+-+

-0.60

mons

-0.59

POSITIVE LOGITS

hips

1.45

pace

1.10

hift

1.07

poons

1.03

pring

1.02

hip

1.00

hin

0.98

pheus

0.95

mith

0.95

poon

0.93

Activations Density 0.021%

words related to a specific name, potentially "Damon."

references to "ors," possibly indicating comparisons or alternatives in various contexts

No Comments

No Known Activations