INDEX

Explanations

spots or locations within a context, possibly related to placement or availability

oai_token-act-pair · gpt-3.5-turbo

references to specific locations or spots

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

yss

-0.76

issance

-0.76

externalActionCode

-0.67

 consortium

-0.65

RR

-0.64

annis

-0.64

ISE

-0.62

 backbone

-0.62

 restraint

-0.62

IFE

-0.61

POSITIVE LOGITS

pots

1.35

 spots

1.15

lights

1.13

hips

1.02

bos

0.90

lot

0.88

pread

0.88

hot

0.88

hots

0.84

coins

0.83

Activations Density 0.007%

spots or locations within a context, possibly related to placement or availability

references to specific locations or spots

No Comments

No Known Activations