INDEX

Explanations

instances where learning is mentioned, particularly from mistakes or experiences

oai_token-act-pair · gpt-3.5-turbo

instances of learning from past experiences and mistakes

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

å¥³

-0.96

 proport

-0.81

itled

-0.80

çĶ°

-0.73

umbered

-0.71

hesion

-0.70

 Goods

-0.70

xus

-0.69

merce

-0.68

itle

-0.68

POSITIVE LOGITS

 experience

1.26

 experiences

1.20

 lessons

1.10

 mistakes

1.04

 history

1.01

 mentors

1.00

 predecessors

0.99

 afar

0.98

exper

0.97

 firsthand

0.96

Activations Density 0.185%

instances where learning is mentioned, particularly from mistakes or experiences

instances of learning from past experiences and mistakes

No Comments

No Known Activations