INDEX

Explanations

short acronyms representing organizations or services

oai_token-act-pair · gpt-3.5-turbo

instances of the abbreviation "LE"

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

ilities

-0.93

ãĥ¼ãĥĨãĤ£

-0.93

ãĤ±

-0.84

ãĤ¢ãĥ«

-0.83

Downloadha

-0.82

raints

-0.80

ãĥĥ

-0.79

udeb

-0.77

Ì¶

-0.76

papers

-0.75

POSITIVE LOGITS

VEL

1.15

IGH

1.09

VEN

0.91

LE

0.88

PLAY

0.86

OVER

0.86

ASE

0.85

VAL

0.85

TE

0.84

LOAD

0.84

Activations Density 0.011%

short acronyms representing organizations or services

instances of the abbreviation "LE"

No Comments

No Known Activations