INDEX

Explanations

mentions of "memory"

oai_token-act-pair · gpt-3.5-turbo

mentions of memory-related concepts

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

ASC

-0.75

 Highlander

-0.72

 foremost

-0.71

 boiling

-0.69

 Nordic

-0.67

 Icelandic

-0.67

 Ducks

-0.66

RAFT

-0.66

 unequal

-0.66

 Palmer

-0.63

POSITIVE LOGITS

oir

1.23

phis

1.17

eor

1.13

orial

1.12

oleon

1.11

oire

1.09

mem

1.08

cript

1.06

pty

1.06

elong

1.05

Activations Density 0.011%

mentions of "memory"

mentions of memory-related concepts

No Comments

No Known Activations