INDEX

Explanations

instances where the word "memory" is mentioned, with a preference for those with a strong emotional or impactful association

oai_token-act-pair · gpt-3.5-turbo

references to memory and the act of remembering

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

retty

-0.73

ahime

-0.72

ModLoader

-0.71

ats

-0.69

wered

-0.69

irin

-0.68

ldon

-0.67

 Galile

-0.67

elsius

-0.66

annis

-0.66

POSITIVE LOGITS

 Memories

1.06

remember

1.04

 remembrance

0.97

 recollection

0.93

 memories

0.91

 remembered

0.88

 remem

0.87

abilia

0.85

 memory

0.84

 retrieval

0.84

Activations Density 0.032%

instances where the word "memory" is mentioned, with a preference for those with a strong emotional or impactful association

references to memory and the act of remembering

No Comments

No Known Activations