INDEX

Explanations

phrases related to learning from mistakes or experiences

oai_token-act-pair · gpt-3.5-turbo

phrases related to 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 @ 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

å¥³

-1.04

xus

-0.87

 proport

-0.78

interstitial

-0.73

ãĥ

-0.72

$$$$

-0.71

inence

-0.70

unless

-0.70

âĺ

-0.70

estate

-0.70

POSITIVE LOGITS

 experience

1.16

 lessons

1.14

 mistakes

1.11

 experiences

1.10

 predecessors

1.07

 past

1.02

 previous

0.99

 successes

0.97

 afar

0.96

 studying

0.96

Activations Density 0.154%

phrases related to learning from mistakes or experiences

phrases related to learning from past experiences and mistakes

No Comments

No Known Activations