INDEX

Explanations

phrases indicating some form of impact or effect

oai_token-act-pair · gpt-3.5-turbo

phrases indicating the concept of leaving an impact or legacy

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

GW

-0.69

uits

-0.67

 hysteria

-0.65

scl

-0.63

 shenan

-0.61

rius

-0.61

ayan

-0.61

GET

-0.60

ilation

-0.60

Attempts

-0.60

POSITIVE LOGITS

 footprints

1.23

 untouched

1.11

 undone

1.10

 residue

1.09

 behind

1.08

 imprint

1.07

 unfinished

1.07

 impression

1.07

 mark

1.04

 scars

1.02

Activations Density 0.144%

phrases indicating some form of impact or effect

phrases indicating the concept of leaving an impact or legacy

No Comments

No Known Activations