INDEX

Explanations

They are looking for statements related to observations or conclusions

oai_token-act-pair · gpt-3.5-turbo

statements that introduce or assert a point or observation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

ickets

-0.91

pees

-0.87

å§«

-0.81

acers

-0.80

tones

-0.79

aths

-0.77

endants

-0.76

aughtered

-0.75

adle

-0.74

aughed

-0.74

POSITIVE LOGITS

 phenomenon

1.11

 begs

1.09

 article

1.07

 imbalance

1.05

 dilemma

1.02

 tendency

1.01

 includes

1.00

 infographic

1.00

 essay

0.99

 applies

0.99

Activations Density 0.154%

They are looking for statements related to observations or conclusions

statements that introduce or assert a point or observation

No Comments

No Known Activations