INDEX

Explanations

words related to performance evaluation and comparison

oai_token-act-pair · gpt-3.5-turbo

mentions of benchmark-related concepts

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

agan

-0.88

oles

-0.77

Else

-0.76

oho

-0.73

frey

-0.73

aina

-0.72

gren

-0.72

artney

-0.71

kn

-0.71

yne

-0.71

POSITIVE LOGITS

 benchmark

1.00

 benchmarks

0.98

marks

0.77

 guiActiveUnfocused

0.70

 scores

0.70

mark

0.69

 score

0.69

 indices

0.68

 suite

0.66

osterone

0.66

Activations Density 0.007%

words related to performance evaluation and comparison

mentions of benchmark-related concepts

No Comments

No Known Activations