INDEX

Explanations

phrases related to political events or geographical locations

oai_token-act-pair · gpt-3.5-turbo

the repeated occurrence of the substring 'ris'

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

è¦ļéĨĴ

-0.85

erm

-0.78

erest

-0.77

ļé

-0.71

resso

-0.69

ishers

-0.66

 shaft

-0.66

lished

-0.66

 EntityItem

-0.62

nesium

-0.62

POSITIVE LOGITS

sey

0.92

py

0.87

pect

0.86

ques

0.85

terday

0.84

coe

0.81

cano

0.79

anchez

0.76

bane

0.75

pell

0.75

Activations Density 0.014%

phrases related to political events or geographical locations

the repeated occurrence of the substring 'ris'

No Comments

No Known Activations