INDEX

Explanations

comparative relationships such as "this or that"

oai_token-act-pair · gpt-3.5-turbo

comparisons or evaluations involving competing options or entities

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

ãĤ¤ãĥĪ

-0.73

 Arrows

-0.71

 scenes

-0.66

enegger

-0.66

ships

-0.64

 faculties

-0.63

ens

-0.61

mercial

-0.61

CLASSIFIED

-0.61

ãĥ¼

-0.59

POSITIVE LOGITS

 apiece

0.68

Lago

0.67

Override

0.67

onite

0.66

 staffer

0.65

rio

0.64

 subp

0.64

per

0.63

ahoo

0.63

icka

0.62

Activations Density 0.310%

comparative relationships such as "this or that"

comparisons or evaluations involving competing options or entities

No Comments

No Known Activations