INDEX

Explanations

acronyms and abbreviations with the specific acronym "HT"

oai_token-act-pair · gpt-3.5-turbo

references to "HT" and related contextual elements

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

 Rite

-0.75

zzi

-0.74

 payable

-0.74

imore

-0.72

Ã±o

-0.72

 Scal

-0.71

oster

-0.67

uum

-0.66

ographer

-0.66

ously

-0.64

POSITIVE LOGITS

dit

0.88

ership

0.81

sa

0.80

chell

0.76

IBLE

0.75

meet

0.73

RON

0.69

hig

0.66

Arcade

0.66

ford

0.65

Activations Density 0.052%

acronyms and abbreviations with the specific acronym "HT"

references to "HT" and related contextual elements

No Comments

No Known Activations