INDEX

Explanations

log messages or commands in a coding context

oai_token-act-pair · gpt-3.5-turbo

brackets and square bracket notations often used in programming or data structures

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 intellig

-0.79

 Boll

-0.67

 dispers

-0.65

 intermitt

-0.65

 morp

-0.63

 occasionally

-0.62

 Explos

-0.62

 vanishing

-0.62

otted

-0.62

 laun

-0.61

POSITIVE LOGITS

actionDate

1.36

?]

1.36

email

1.25

!]

1.25

](

1.20

REDACTED

1.17

%]

1.14

...]

1.13

=]

1.09

][

1.07

Activations Density 0.036%

log messages or commands in a coding context

brackets and square bracket notations often used in programming or data structures

No Comments

No Known Activations