INDEX

Explanations

text related to computing and technical instructions

oai_token-act-pair · gpt-3.5-turbo

indications of processes or functions that relate to division and segmentation in various contexts, often represented by the "/" symbol

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 Grimes

-0.78

 Giov

-0.75

 Carlton

-0.74

 Angelo

-0.72

 Canter

-0.71

 Tasman

-0.71

Wyn

-0.71

 Kern

-0.71

 Chambers

-0.70

 Tornado

-0.69

POSITIVE LOGITS

etc

1.57

whatever

1.35

Flickr

1.13

pmwiki

1.08

workshop

1.06

photos

1.00

legal

1.00

coll

0.98

non

0.97

oops

0.96

Activations Density 0.048%

text related to computing and technical instructions

indications of processes or functions that relate to division and segmentation in various contexts, often represented by the "/" symbol

No Comments

No Known Activations