INDEX

Explanations

phrases related to explaining how something operates or works

oai_token-act-pair · gpt-3.5-turbo

phrases that describe functionality or the workings of various systems or mechanisms

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

gart

-0.81

voy

-0.69

imar

-0.68

TERN

-0.68

Bir

-0.67

miss

-0.65

moon

-0.64

ark

-0.61

intage

-0.60

ãģķ

-0.60

POSITIVE LOGITS

 proced

0.72

 unfolded

0.68

:[

0.67

 nutshell

0.65

 differs

0.65

 unfold

0.65

iquette

0.62

 Different

0.62

 unfolding

0.62

 shaping

0.62

Activations Density 0.097%

phrases related to explaining how something operates or works

phrases that describe functionality or the workings of various systems or mechanisms

No Comments

No Known Activations