INDEX

Explanations

phrases related to returning or revisiting something

oai_token-act-pair · gpt-3.5-turbo

instances of the word "back" used in various contexts, particularly related to return or recurrence

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

 understatement

-0.67

rez

-0.64

icipated

-0.63

aspx

-0.62

orst

-0.61

"]=>

-0.61

 Telescope

-0.61

alion

-0.61

esan

-0.60

anto

-0.60

POSITIVE LOGITS

doors

0.94

packs

0.88

door

0.86

stories

0.83

seat

0.82

tracking

0.82

wards

0.82

home

0.80

GROUND

0.80

fires

0.78

Activations Density 0.043%

phrases related to returning or revisiting something

instances of the word "back" used in various contexts, particularly related to return or recurrence

No Comments

No Known Activations