INDEX

Explanations

references to messages

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 @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

-0.80

aughs

-0.74

itals

-0.73

arte

-0.69

erenn

-0.68

emale

-0.67

ONSORED

-0.66

urses

-0.66

pmwiki

-0.66

rowd

-0.66

POSITIVE LOGITS

 messages

0.98

 goodbye

0.91

board

0.91

box

0.89

message

0.88

 message

0.87

boxes

0.86

 Messages

0.86

FontSize

0.84

 sent

0.83

Activations Density 0.032%

references to messages

references to messages or communications

No Comments

No Known Activations