INDEX

Explanations

programming-related terms like types, functions, parameters, and file operations

oai_token-act-pair · gpt-3.5-turbo

concepts related to object creation and programming principles

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 Dame

-0.72

Mug

-0.70

 Evening

-0.69

 UKIP

-0.68

 Reeves

-0.67

 Griffith

-0.65

 Bennett

-0.65

 Brock

-0.64

 Trend

-0.63

 celebrity

-0.62

POSITIVE LOGITS

 inheritance

1.06

 abstraction

1.04

 interpol

1.03

 initialization

1.02

 annotations

1.01

Iterator

1.01

 syntax

1.01

 liter

0.99

 semantics

0.98

 parsing

0.98

Activations Density 0.191%

programming-related terms like types, functions, parameters, and file operations

concepts related to object creation and programming principles

No Comments

No Known Activations