INDEX

Explanations

phrases related to specific codes or acronyms, particularly those related to AA

oai_token-act-pair · gpt-3.5-turbo

instances of the abbreviation "AA"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 Jenner

-0.78

icles

-0.74

icle

-0.72

 Canaver

-0.72

ophers

-0.72

pheus

-0.71

opher

-0.68

ledge

-0.67

sticks

-0.66

points

-0.66

POSITIVE LOGITS

BILITY

0.97

BILITIES

0.93

BLE

0.91

SPA

0.90

ZZ

0.88

UGH

0.87

qua

0.87

HA

0.86

HAHAHAHA

0.86

ULTS

0.86

Activations Density 0.008%

phrases related to specific codes or acronyms, particularly those related to AA

instances of the abbreviation "AA"

No Comments

No Known Activations