INDEX

Explanations

numbered lists or bullet points

oai_token-act-pair · gpt-3.5-turbo

structured information, particularly lists or items, such as features or contents in a release note

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

imaru

-0.81

hement

-0.75

 emancipation

-0.72

tsky

-0.70

 intervened

-0.69

 inacc

-0.68

 paralysis

-0.67

 diseng

-0.65

 Saras

-0.65

 midway

-0.64

POSITIVE LOGITS

Original

0.83

thumbnails

0.82

INST

0.80

Beta

0.78

Website

0.78

âĹı

0.77

Available

0.76

rawdownloadcloneembedreportprint

0.76

OUNT

0.75

âľ

0.75

Activations Density 0.098%

numbered lists or bullet points

structured information, particularly lists or items, such as features or contents in a release note

No Comments

No Known Activations