INDEX

Explanations

names of characters or people

oai_token-act-pair · gpt-3.5-turbo

characters and their actions within a narrative context

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Repeat

-0.79

IED

-0.77

interstitial

-0.77

Depth

-0.72

PLIED

-0.72

NUM

-0.71

orted

-0.69

vered

-0.68

TPPStreamerBot

-0.67

IFIED

-0.67

POSITIVE LOGITS

 discovers

1.70

 learns

1.61

 escapes

1.51

 convin

1.49

 decides

1.48

 confronts

1.48

 realizes

1.43

 wakes

1.38

 tries

1.35

 finds

1.34

Activations Density 0.255%

names of characters or people

characters and their actions within a narrative context

No Comments

No Known Activations