INDEX

Explanations

corrections or errors in text

oai_token-act-pair · gpt-3.5-turbo

terms related to correctness and error in information or statements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 Phant

-0.73

ota

-0.69

lust

-0.69

 Skies

-0.68

 spirit

-0.66

wolves

-0.63

 spirits

-0.62

ofi

-0.60

 enjoy

-0.60

 aura

-0.60

POSITIVE LOGITS

correct

3.35

Correct

3.12

 Correct

2.68

Correction

2.44

 Wrong

2.02

 Correction

2.00

 corrections

1.94

wrong

1.93

 incorrect

1.82

orrect

1.82

Activations Density 0.025%

corrections or errors in text

terms related to correctness and error in information or statements

No Comments

No Known Activations