INDEX

Explanations

coordinated phrases, possibly related to a sequence or grouping

oai_token-act-pair · gpt-3.5-turbo

conjunctions and expressions indicating continuation or emphasis in dialogue

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

AME

-0.80

ige

-0.80

activity

-0.70

actionDate

-0.69

 âĢİ

-0.69

ajor

-0.68

Pg

-0.67

SPA

-0.67

iche

-0.67

athetic

-0.66

POSITIVE LOGITS

 congr

1.01

romeda

0.93

 optionally

0.88

vo

0.82

 secondly

0.81

 congratulations

0.78

 beware

0.75

 importantly

0.75

 incidentally

0.71

rew

0.71

Activations Density 0.184%

coordinated phrases, possibly related to a sequence or grouping

conjunctions and expressions indicating continuation or emphasis in dialogue

No Comments

No Known Activations