INDEX

Explanations

verbs related to actions or decisions, particularly involving confirmations or expressions

oai_token-act-pair · gpt-3.5-turbo

references to political decisions and controversies

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 unlucky

-0.73

 sucks

-0.71

Veh

-0.66

BET

-0.65

 lousy

-0.64

Tes

-0.63

soDeliveryDate

-0.63

 haun

-0.63

 Experiment

-0.62

 Failed

-0.61

POSITIVE LOGITS

 publicly

1.22

 formally

1.16

 specifics

1.09

 explicitly

1.09

 divul

1.03

 commented

1.03

 officially

1.00

 comment

1.00

osponsors

0.98

 disclosed

0.96

Activations Density 0.245%

verbs related to actions or decisions, particularly involving confirmations or expressions

references to political decisions and controversies

No Comments

No Known Activations