INDEX

Explanations

words related to specific details or elements in technical contexts

oai_token-act-pair · gpt-3.5-turbo

keywords associated with specific entities or concepts, such as links, food items, and groups

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

Sensor

-0.69

_-

-0.66

consuming

-0.65

icipated

-0.64

aver

-0.63

orthern

-0.63

Ïī

-0.62

erd

-0.62

aring

-0.61

nox

-0.61

POSITIVE LOGITS

 syndrome

0.83

 path

0.65

Qi

0.64

ilib

0.59

æ

0.59

 Syndrome

0.58

 Pants

0.57

Tai

0.57

odies

0.56

(>

0.55

Activations Density 0.615%

words related to specific details or elements in technical contexts

keywords associated with specific entities or concepts, such as links, food items, and groups

No Comments

No Known Activations