INDEX

Explanations

proper nouns and terms related to characters, especially in creative works like films, games, and books

oai_token-act-pair · gpt-3.5-turbo

references to characters in stories or narratives

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

Accessory

-0.71

Effective

-0.64

LOCK

-0.64

ribution

-0.64

ctr

-0.62

 Kremlin

-0.62

IED

-0.60

 unilateral

-0.60

 Tant

-0.60

 Wass

-0.59

POSITIVE LOGITS

acters

1.44

istically

0.90

hips

0.89

istics

0.82

hip

0.81

poons

0.78

 Characters

0.77

 characters

0.76

 arcs

0.76

 inhab

0.74

Activations Density 0.021%

proper nouns and terms related to characters, especially in creative works like films, games, and books

references to characters in stories or narratives

No Comments

No Known Activations