INDEX

Explanations

mentions of tools or tool-related activities

oai_token-act-pair · gpt-3.5-turbo

references to tools and related concepts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 Sons

-0.89

 sons

-0.64

 Qiao

-0.64

otos

-0.61

 mosqu

-0.61

egal

-0.60

oul

-0.60

ategory

-0.58

 adolesc

-0.58

ilipp

-0.58

POSITIVE LOGITS

kit

1.50

tips

1.50

tip

1.14

sonian

1.10

belt

1.06

chain

1.01

assisted

0.95

chains

0.95

box

0.93

 tools

0.92

Activations Density 0.036%

mentions of tools or tool-related activities

references to tools and related concepts

No Comments

No Known Activations