INDEX

Explanations

phrases indicating disagreement or objection

oai_token-act-pair · gpt-3.5-turbo

phrases connected to government actions and personal rights issues

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

raft

-0.68

erenn

-0.63

 Kinnikuman

-0.58

 Gamble

-0.56

mma

-0.56

issors

-0.53

 ----------------

-0.53

phal

-0.52

eve

-0.52

minus

-0.51

POSITIVE LOGITS

 accordingly

0.77

 oneself

0.71

 thereto

0.70

 them

0.68

 themselves

0.68

 thereafter

0.67

 thereof

0.67

)?

0.63

 exceptions

0.62

 nonetheless

0.60

Activations Density 1.630%

phrases indicating disagreement or objection

phrases connected to government actions and personal rights issues

No Comments

No Known Activations