INDEX

Explanations

words related to providing additional information or directing attention to specific details

oai_token-act-pair · gpt-3.5-turbo

phrases or references to locations and repeated structures in text

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

thood

-0.70

Malley

-0.67

udence

-0.66

olation

-0.60

omy

-0.59

ims

-0.59

imi

-0.59

agi

-0.58

piration

-0.58

Buyable

-0.58

POSITIVE LOGITS

 Nieto

0.81

 DISTRICT

0.68

swick

0.68

 dotted

0.66

nesday

0.64

 dots

0.64

OULD

0.63

å¥³

0.62

ican

0.61

Zar

0.59

Activations Density 0.198%

words related to providing additional information or directing attention to specific details

phrases or references to locations and repeated structures in text

No Comments

No Known Activations