INDEX

Explanations

names or terms related to specific individuals, potentially as part of a story or narrative

oai_token-act-pair · gpt-3.5-turbo

proper nouns, especially names associated with individuals and places

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

icles

-0.78

ctic

-0.74

ioch

-0.69

-0.66

icle

-0.66

iest

-0.66

iers

-0.65

ushed

-0.62

quet

-0.61

ushing

-0.60

POSITIVE LOGITS

otive

0.85

wich

0.83

otion

0.81

etheus

0.79

hole

0.78

uria

0.77

ajor

0.77

rade

0.76

insky

0.75

clair

0.74

Activations Density 0.098%

names or terms related to specific individuals, potentially as part of a story or narrative

proper nouns, especially names associated with individuals and places

No Comments

No Known Activations