INDEX

Explanations

numerical ranges such as "one to nine" or "one to eight"

oai_token-act-pair · gpt-3.5-turbo

concepts related to one-off events or experiences

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

Hoo

-0.67

âĸ¬âĸ¬

-0.61

 GOODMAN

-0.59

âĶģ

-0.59

 Stra

-0.58

 resil

-0.58

ITED

-0.57

Downloadha

-0.57

 laun

-0.56

IBLE

-0.56

POSITIVE LOGITS

sided

1.36

dimensional

1.34

legged

1.14

third

1.14

handed

1.12

minute

1.11

eyed

1.08

liners

1.08

year

1.07

piece

1.07

Activations Density 0.044%

numerical ranges such as "one to nine" or "one to eight"

concepts related to one-off events or experiences

No Comments

No Known Activations