INDEX

Explanations

terms related to benchmarking

oai_token-act-pair · gpt-3.5-turbo

terms related to benchmarks across various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

agan

-0.81

oho

-0.78

otin

-0.74

Else

-0.74

kn

-0.74

urses

-0.74

por

-0.73

oles

-0.72

thel

-0.71

cha

-0.70

POSITIVE LOGITS

 benchmarks

1.19

 benchmark

1.19

 guiActiveUnfocused

0.82

 indices

0.79

marks

0.76

bench

0.76

 metrics

0.74

 glim

0.70

mark

0.68

 Bench

0.68

Activations Density 0.010%

terms related to benchmarking

terms related to benchmarks across various contexts

No Comments

No Known Activations