INDEX

Explanations

strings indicating errors and instructions to correct them, likely related to subscriptions or forms

oai_token-act-pair · gpt-3.5-turbo

punctuations and formatting errors 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

 brut

-0.76

 manif

-0.75

oun

-0.74

anus

-0.74

bos

-0.74

zik

-0.72

oshenko

-0.72

 induct

-0.71

cod

-0.70

ouf

-0.70

POSITIVE LOGITS

 Please

1.04

php

0.96

 However

0.94

dll

0.93

google

0.92

push

0.92

html

0.91

jpg

0.88

 Although

0.88

 Typically

0.87

Activations Density 0.114%

strings indicating errors and instructions to correct them, likely related to subscriptions or forms

punctuations and formatting errors in text

No Comments

No Known Activations