INDEX

Explanations

message-related terms

oai_token-act-pair · gpt-3.5-turbo

references to messages or communications

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

engeance

-0.76

 Sketch

-0.72

abama

-0.71

aughs

-0.70

emale

-0.69

pmwiki

-0.69

itals

-0.67

itect

-0.66

erenn

-0.66

nec

-0.66

POSITIVE LOGITS

board

1.08

 messages

1.06

 message

1.06

 boards

1.02

boards

1.02

boxes

0.97

message

0.96

 sent

0.95

 conveyed

0.95

 sender

0.92

Activations Density 0.035%

message-related terms

references to messages or communications

No Comments

No Known Activations