INDEX

Explanations

phrases related to technical instructions or programming code

oai_token-act-pair · gpt-3.5-turbo

identifiers and parameters related to commands and functionalities in a technical or programming context

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

ensis

-0.73

ilogy

-0.58

omore

-0.58

hement

-0.56

cli

-0.56

etheless

-0.56

lease

-0.55

hem

-0.54

ordering

-0.54

perty

-0.54

POSITIVE LOGITS

 Accountability

0.70

Zero

0.68

 Subtle

0.67

 Country

0.67

 Commons

0.66

Own

0.66

 Execution

0.65

 Majesty

0.65

Bok

0.65

 Rewards

0.65

Activations Density 1.124%

phrases related to technical instructions or programming code

identifiers and parameters related to commands and functionalities in a technical or programming context

No Comments

No Known Activations