INDEX

Explanations

actions or processes that lead to an outcome, improvement, or change

oai_token-act-pair · gpt-3.5-turbo

actions that signify change or effect in various contexts

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

nex

-0.70

volent

-0.65

conn

-0.63

zo

-0.63

street

-0.59

shore

-0.58

 Antar

-0.58

SW

-0.58

sw

-0.58

squ

-0.57

POSITIVE LOGITS

ettings

0.99

hift

0.95

paces

0.95

ometimes

0.91

hement

0.79

hirt

0.79

ilver

0.76

creen

0.74

pace

0.74

heet

0.71

Activations Density 0.462%

actions or processes that lead to an outcome, improvement, or change

actions that signify change or effect in various contexts

No Comments

No Known Activations