INDEX

Explanations

phrases about common human experiences or traits with a focus on uniqueness

oai_token-act-pair · gpt-3.5-turbo

phrases indicating possession or existence related to individuals

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

inth

-0.82

Newsletter

-0.70

ound

-0.69

etting

-0.69

live

-0.68

edia

-0.68

QUI

-0.68

ãĤº

-0.66

sole

-0.66

cise

-0.65

POSITIVE LOGITS

 flaws

0.97

 biases

0.88

 quirks

0.87

 weaknesses

0.84

 faults

0.83

 differing

0.83

 varying

0.83

 precon

0.82

 strengths

0.82

 undergone

0.78

Activations Density 0.183%

phrases about common human experiences or traits with a focus on uniqueness

phrases indicating possession or existence related to individuals

No Comments

No Known Activations