INDEX

Explanations

words related to rewards and punishments

oai_token-act-pair · gpt-3.5-turbo

terms related to rewards and punishments

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

jet

-0.75

Champ

-0.74

cre

-0.69

walker

-0.68

creation

-0.67

coat

-0.66

 Takeru

-0.66

 Blazers

-0.66

 Surf

-0.66

 Liber

-0.66

POSITIVE LOGITS

 rewarded

1.08

 punish

1.04

 punished

0.93

ricular

0.93

 rewarding

0.93

 punishing

0.87

 incentiv

0.85

 harshly

0.80

 showc

0.78

 reward

0.76

Activations Density 0.014%

words related to rewards and punishments

terms related to rewards and punishments

No Comments

No Known Activations