INDEX

Explanations

phrases related to a particular technology or product, likely containing the abbreviation "GA"

oai_token-act-pair · gpt-3.5-turbo

mentions of "GA" and related organization or project identifiers

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

alities

-0.77

ness

-0.74

meyer

-0.73

names

-0.67

 Wond

-0.67

tie

-0.66

ties

-0.65

hood

-0.65

 surn

-0.64

gie

-0.63

POSITIVE LOGITS

UGE

1.12

ZA

1.00

ZI

0.98

 TOUR

0.91

ULTS

0.90

BA

0.88

HHHH

0.87

ZE

0.85

RECT

0.85

ESA

0.84

Activations Density 0.011%

phrases related to a particular technology or product, likely containing the abbreviation "GA"

mentions of "GA" and related organization or project identifiers

No Comments

No Known Activations