INDEX

Explanations

phrases related to events happening at a later time

oai_token-act-pair · gpt-3.5-turbo

phrases indicating later revelations or statements in the 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

 Container

-0.70

Plot

-0.69

minecraft

-0.68

afety

-0.67

yet

-0.67

cius

-0.66

 encyclopedia

-0.62

 Mahjong

-0.61

anium

-0.61

 Simpl

-0.60

POSITIVE LOGITS

 recons

0.87

 forg

0.80

 wiser

0.79

livion

0.72

 regretted

0.72

 relent

0.71

pez

0.71

 realise

0.67

 sidx

0.64

}\

0.63

Activations Density 0.221%

phrases related to events happening at a later time

phrases indicating later revelations or statements in the text

No Comments

No Known Activations